1 Answer
- Newest
- Most votes
- Most comments
0
If you call the PutEvents API without a userId to send in new interactions for an anonymous user, you can call the GetRecommendations API after 1-2 seconds with the sessionId as the userId to retrieve recommendations that are personalized to the session. Of course, the sessionId from the PutEvents call needs to be the same value in the GetRecommendations API call.
If your PutEvents call(s) have a sessionId and userId specified, you must use the userId as the userId in the GetRecommendations call.
answered 2 years ago
Relevant content
- asked 5 months ago
- asked 2 years ago
- AWS OFFICIALUpdated 7 months ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 5 months ago
When we do a bulk upload from a csv (dataset import job), with interactions having
session_id
s but nouser_id
s, is it supposed to be the same way? Or is this functionality dependent upon using thePutEvents
functionality? Thanks for your quick responseReal-time recommendations (i.e., where recommendations change in (near) real-time based on new interactions) requires that a campaign (or recommender) be used and new interactions are streamed in via the PutEvents API. This also only applies to specific recipes. Check the docs for details. https://docs.aws.amazon.com/personalize/latest/dg/recording-events.html
New interactions uploaded via import jobs (bulk or incremental imports), will only impact recommendations for known users after the model is fully retrained.